J. Am. Chem. Soc. 1999, 121, 5093-5094
5093
Decomposition of Sodium Trioxodinitrate (Angeli’s Salt) To Hydroxyl Radical: An ESR Spin-Trapping Study Detcho A. Stoyanovsky,*,† Robert Clancy,‡ and Arthur I. Cederbaum† Department of Biochemistry Mount Sinai School of Medicine New York, New York 10029 Department of Rheumatology Hospital for Joint Diseases, New York UniVersity New York, New York ReceiVed February 1, 1999 ReVised Manuscript ReceiVed April 20, 1999 There is much interest on the preponderant redox form of nitric oxide under physiological conditions. Despite the controversy whether nitric oxide synthase catalyzes the synthesis of nitroxyl anion (NO-) or nitric oxide (NO•),1,2 various mechanisms for production of NO- that may operate under physiological conditions have been reported. Superoxide dismutase can convert NO• to NO- 3 and S-nitrosothiols, a class of compounds implicated as playing key roles in a variety of physiological processes, can react with thiols to generate NO-.4 In model studies aimed to mimic the biochemistry of NO-, sodium trioxodinitrate (Na2N2O3, Angeli’s salt, AS) is often used as a specific NO- donor.1,5 The hydrolysis of AS is a complex process that may trigger reactions of alkylation, condensation, and oxidation6 by mechanism(s) not well understood and which cannot be correlated only with the hydrolysis of AS to NO-. In this report, we present ESR spintrapping data which suggest that AS undergoes a pH-dependent decomposition to hydroxyl radical. The Angeli-Rimini test for aldehydes is based on the ASdependent formation of hydroxamic acids.7 The reaction is mechanistically believed to proceed via addition of the hydroxylamine nitrogen of the trioxodinitrate anion to the carbonyl carbon atom of the aldehyde; a subsequent release of NO2- and a tautomerization of the CH-NdO structure of the resulting R-nitroso alcohol forms the corresponding hydroxamic acids, which can be detected spectrophotometrically after chelation with FeCl3.6,7 Depending on the degree of protonation, the stability of AS in aqueous solutions follows the sequence N2O32- f HN2O3f H2N2O3 (pK1 ) 3.0 and pK2 ) 9.358). AS is relatively stable in alkaline solutions; its rate of decomposition within the pH interval 3.5-8.5 is rapid and [H+]-independent, proceeding via an intermediate formation of NO-, and NO2- and N2O as end products. At lower pH values, the decomposition rate increases * Author to whom correspondence should be sent. Telephone: 212-2417285. Fax: 212-996-7214. † Mt. Sinai School of Medicine. ‡ Hospital for Joint Diseases. (1) Schmidt, H. H.; Hofmann, H.; Schindler, U.; Shutenko, Z. S.; Cunningham, D. D.; Feelisch, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 14492-14497. (2) Xia, Y.; Zweier, J. L. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 1270512710. (3) Murphy, M. E.; Sies, H. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1086010864. (4) Wong, P. S.; Hyun, J.; Fukuto, J. M.; Shirota, F. N.; DeMaster, E. G.; Shoeman, D. W.; Nagasawa, H. T. Biochemistry 1998, 37, 5362-5371. (5) Ohshima, H.; Yoshie, Y.; Auriol, S.; Gilibert, I. Free Radical Biol. Med. 1998, 25, 1057-1065. DeMaster, E. G.; Redfern, B.; Nagasawa, H. T. Biochem. Pharmacol. 1998, 55, 2007-2015. Vanuffelen, B. E.; Van Der Zee, J.; De Koster, B. M.; Vansteveninck, J.; Elferink, J. G. Biochem. J. 1998, 330, 719-722. (6) Smith, P. A. S.; Hein, G. E. J. Am. Chem. Soc. 1960, 82, 5731-5740. (7) Angeli, A. Gazz. Chim. Ital. 1896, 26, 17. (8) Hughes, M. N.; Wimbledon, P. E. J. Chem. Soc., Dalton 1976, 703707.
Figure 1. ESR spectra of DMPO nitroxides formed in a solution of AS. ESR measurements and spectra simulations were performed as described in ref 11. DMPO was used at a concentration of 0.1 M in phosphate buffer (pH 7.4, 25 °C). The hyperfine splitting constants (in G) used for simulation of the spectra of the DMPO/•OH, DMPO/HER, and DMPO/ •CH nitroxides were as follows: (A ) 15.0; A ) 15.0), (A ) 15.8; 3 N H N AH ) 22.8) and (AN ) 16.3; AH ) 23.2), respectively (11). AS was synthesized as described in (6). All ESR spectra were recorded after 2 min of incubation of the reaction solutions. Spectra 1, no addition; spectra 2, plus AS (0.1 mM; a stock solution of AS was prepared in 0.2 M NaOH); spectra 3, computer simulation of the ESR spectra of DMPO/ •OH; spectra 4, plus ethanol (0.5 M) and AS; spectra 5, ethanol and AS were added after deoxygenation of all solutions with N2 for 45 min; spectra 6, plus DMSO (0.3 M) and AS; trace 7 represents computer simulation of the ESR spectra of DMPO/HER (solid lines) and DMPO/ .CH (dashed lines). The ESR spectra in repetitive experiments did not 3 differ more than 10% (n g 3).
with increasing acidity with production of NO•.8 While the structure of AS has been determined to be [ONdNO2]2-, there is uncertainty over the exact mechanism(s) of its decomposition in aqueous solutions. Studies with [15N2O3]2- have suggested that a nitrite anion is released via NdN cleavage of the nitrogen that is bound to two oxygen atoms.9 At pH > 4 AS is decomposed, depending on the site of protonation, to either HNO (pKa ) 4.7) and NO2-, or to NO- and HNO2.8,9 AS can be stabilized by sodium nitrite, presumably due to a recombination of HNO with NO2- to HN2O3-;8 at pH < 4, however, sodium nitrite accelerates the decomposition of AS via a HNO2-dependent consumption of H2N2O3.8 The rate of AS decomposition at pH values less than 3 decreased in the presence of ethanol, suggesting that the hydrolytic process may proceed via a HO•-dependent rate-limiting step.8 Thermal decomposition of AS in aqueous solutions generates peroxynitrite anion (ONOO-) via the formation of singlet NO-; upon relaxation of singlet-state NO-, a triplet state NO- is formed that reacts with oxygen to form ONOO-.10 The physiological effects of NO- are associated with its potential to selectively oxidize low-molecular weight and protein thiols via a nonradical mechanism, and to form complexes with heme proteins.1 However, in ESR spin-trapping experiments with (9) Hendrickson, D. N.; Jolly, W. L. Inorg. Chem. 1969, 8, 693-700. (10) Donald, C. E.; Hughes, M. N.; Thompson, J. M.; Bonner, F. T. Inorg. Chem. 1986, 25, 2676-2677.
10.1021/ja990318v CCC: $18.00 © 1999 American Chemical Society Published on Web 05/14/1999
5094 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Communications to the Editor
Scheme 1
AS-loaded murine macrophage cells, we observed the formation of radical intermediates (data not shown). This led us to study whether the hydrolysis of AS is associated with the generation of HO•. Figure 1.2 depicts the ESR spectra of a cell-free solution of AS in 5,5′-dimethyl-1-pyroline N-oxide (DMPO, 1; Scheme 1)-containing phosphate buffer (pH 7.3). The appearance of a fourline ESR spectra with hyperfine structure (in G) of AN ) 15.0 and AH ) 15.0 allows the assignment of the adduct as that formed by addition of ΗΟ• to DMPO (Scheme 1, 2;).11 Catalase (5002500U/ml) and superoxide dismutase (30-3000 U/ml) did not affect the AS-dependent formation of DMPO/•OH, suggesting that H2O2 and superoxide anion are not involved in this production. A maximal rate of DMPO/•OH production was observed at pH 5 (Figure 2A). The process had a bi-phasic kinetic profile with increasing values over a time period of 10 min. The maximal, steady-state spin concentration of the DMPO/•OH nitroxide was 5.4 µM as determined by double integration of the ESR signal using 4-hydroxyl-1-TEMPO as a standard.12 Ethanol and dimethyl sulfoxide (DMSO) are oxidized by HO• to 1-hydroxyethyl radical (HER11) and methyl radical, respectively;13 the hydrolysis of AS in the presence of DMPO and either ethanol or DMSO produced a typical ESR spectra of the DMPO/ HER or DMPO/•CH3 nitroxide (Scheme 1, 3 and 4, respectively; Figure 1, traces 4 and 5: AN ) 15.8 G, AH ) 22.8 G; trace 6: AN ) 16.3 G, AH ) 23.2 G): AS can generate ONOO- under aerobic conditions which may affect the ESR spin-trapping measurements (ref 14 and the references therein). Removal of oxygen from the reaction solutions, however, did not decrease the ESR spectra of DMPO/HER (trace 5 compared to trace 4), suggesting that the formation of the nitroxide follows a HO•- rather than a ONOO--dependent abstraction of R-hydrogen atom from ethanol. The AS-dependent generation of HO• was further supported by determination of the rate constant for the interaction of HO• with DMPO as described in ref 15; we obtained a value of 3.52 × 109 M-1 s-1, which is in good agreement with the literature values of 2.1-3.4 × 109 M-1 s-1.15 HER has a short lifetime and, after an electron transfer to oxygen, is converted to acetaldehyde. Figure 2B illustrates the accumulation of acetaldehyde in an ethanol-containing solution of AS. It is likely that in this system a considerable part of the generated acetaldehyde reacts with AS to produce acetyl hydroxamic acid (Scheme 2). In support of the latter assumption are the results from the FeCl3 color reaction for hydroxamic acids (Angeli-Rimini test); an addition of FeCl3 into AS-containing solutions of either ethanol or acetaldehyde led to qualitatively equal spectral changes within the visible range (Figure 2C, spectra 2 and 3, respectively). No FeCl3-dependent changes were observed in the absence of ethanol or acetaldehyde (Figure 2C, spectrum 1). (11) Stoyanovsky, D. A.; Wu, D.; Cederbaum, A. I. Free Radical Biol. Med. 1998, 24, 132-138. (12) Morrisett, J. D. Spin Labeling Theory and Applications; Berliner L. J., ed.; Academic Press: Orlando, FL, 1976; pp 272-338. (13) Klein, S. M.; Cohen, G.; Cederbaum, A. I. Biochemistry 1981, 20, 6006-6012. (14) Gatti, R. M.; Alvarez, B.; Vasquez-Vivar, J.; Radi, R.; Augusto, O. Arch. Biochem. Biophys. 1998, 349, 36-48. (15) Morehouse, K. M.; Mason, R. P. J. Biol. Chem. 1988, 263, 12041211. Buettner, G. R.; Motten, A. G.; Hall, R. D.; Chignell, C. F. Photochem. Photobiol. 1986, 44, 5-10. Knecht, K. T.; Thurman, R. G.; Mason, R. P. Arch. Biochem. Biophys. 1993, 303, 339-48.
Figure 2. pH-dependent hydrolysis of AS to hydroxyl radical. A. ESR signal intensity of AS in DMPO-containing phosphate buffer as a function of pH. (b) AS; (O), AS plus NaCl (1 M); (2), AS plus NaNO2 (2 mM). B. GC-monitored accumulation of acetaldehyde in AS (0.3 mM)- and ethanol (0.5 M)-containing phosphate buffer (0.1 M, pH ) 5). GC (Hewlett-Packard 5840A gas chromatograph) measurements were conducted as described in ref 17; (b) plus AS; (O) minus AS. C. Effects of FeCl3 (0.02 mM) on the visible spectra of a solution of AS (0.3 mM, spectra 1), AS plus ethanol (0.5M, spectra 2), and AS plus acetaldehyde (0.01 M, spectra 3) in 0.1 M phosphate buffer (pH ) 5). The reaction solutions were kept for 4 min at 20 °C prior to the addition of FeCl3.
Scheme 2
Scheme 3
Several mechanisms may account for the AS-dependent generation of HO•. It is anticipated that NO- can dimerize to cishyponitrous acid (H2N2O28), which is unstable and decomposes via a chain-reaction mechanism that is most likely initiated by an azo-type homolytic fission (HO-NdN-OH f N2 + 2 HO• 16). The ability of NaCl (but not NaClO4; data not shown) to stimulate the production of DMPO/•OH in AS-containing solutions suggests that H2N2O2 is a potential source of hydroxyl radical (Figure 2A); NaCl has been shown to catalyze the decomposition of H2N2O2.16 Contrary to NaCl, NaNO2 inhibited the AS-dependent production of DMPO/•OH (Figure 2A), presumably by decreasing the concentration of H2N2O2 via displacing the equilibrium between HN2O3-, HNO, and the endogenous NO2- in favor of HN2O3- 8 (Scheme 3): At certain pH values, the AS-dependent generation of HO• may reflect the occurrence of more complex reactions,8 or as a reviewer has pointed out, HO• can be preceded by an electron exchange between HN2O3- and H2N2O3 (or N2O23 and HN2O3-). Independently of the exact mechanism(s), two immediate implications of the AS-dependent generation of HO. may be envisaged: (1) Contrary to NO-, hydroxyl radical is a species that interacts indiscriminately, in a diffusion-controlled manner with low-molecular weight compounds, lipids, and proteins present in cells. In AS solutions, the formation of HO• parallels the formation of NO-, suggesting that the use of Angeli’s salt as a NO- donor in biological systems should be accompanied by appropriate controls for discrimination of NO-- from HO•dependent effects, and (2) AS can convert primary alcohols to the corresponding hydroxamic acids, extending in this way the application of Angeli-Rimini test to the detection of alcohols. Acknowledgment. This work was supported by USPHS Grant AA09460 from The National Institute on Alcohol Abuse and Alcoholism.
JA990318V (16) Buchholz, J. R.; Powell, R. E. J. Am. Chem. Soc. 1965, 87, 23502353. (17) Cederbaum, A. I.; Cohen, G. Methods Enzymol. 1984, 105, 516522.
